Non-Traumatic Intracranial Hemorrhages: Clinical-Statistical Description and Multifactor Analysis of Outcomes

T.N. Galkina

Russian Polenov Neurosurgical Institute Saint Petersburg, Russia

Introduction

Non-traumatic intracranial hemorrhages (NICH), being the most severe manifestation of vascular pathology of the brain, are widespread. They account for 20% of all cerebrovascular pathology and are characterized by high mortality and invalidism [8]. There is further increase of their rate, watched in Russia during the last years against a background of a difficult socio-economic situation [6, 12].

Today there is no unanimous opinion on causes of NICH, a role of risk factors in development of this pathology and factors, effecting outcomes [3, 14]. As for reports on a period of time between hemorrhage, rendering medical care and admission to a hospital, and their effect on NICH outcomes, they are contradictory [4].

There is no conventional system of rendering medical care to patients with NICH in Russia. Patients are admitted not only to medical establishments, specializing in treatment of vascular pathology, but also to multi-field hospitals, which have no neurosurgical departments and facilities for CT-examination and where diagnostic procedures are limited to neurologic examination and lumbar puncture. The result of this situation, existing in Saint Petersburg during the last years, is rather high mortality (55.76-57.45%) in patients with HICH. Thus, it has become a cause of the present study.

The Study Goal

It lay in investigating a structure of NICH, revealing and analyzing factors, which effect outcomes, and working out principles of improvement of medical care in this pathology.

Material and Methods

There were two trends of activity: study of such medical-clinical aspects, as a structure, severity and outcomes of NICH and study of medical-organizational problems, including causes and risk factors of development of this pathology, as well as organization of medical care at a prehospital and hospital stages.

Patients with NICH (682 cases), treated in three hospitals of Saint Petersburg, were subject to complex study.

These hospitals differ from each other in a degree of specialization and material and technical support. The Mariinsky Hospital was regarded to be a hospital of the third, i.e. highest level; it has a neurosurgical department, specializing in treatment of patients with vascular pathology, and CT. The Saint Elizabeth Hospital was referred to a medical establishment of the second level; it has a neurosurgical department, which does not specialize in rendering care to the above patients, and CT. The Pokrovskaya Hospital was considered to be a hospital of the first and lowest level; it has no neurosurgical department and constantly operating CT.

We failed to find estimation of patients' state on the basis of generally accepted classifications (Hunt and Hess Scale, Glasgow Coma Score Scale) in their case reports. Thus, we assessed a degree of consciousness disturbance as an integrative factor, capable of characterizing a patient state best of all, for carrying out multifactor analysis of NICH outcomes.

We were the first to use a new approach to studying a system of rendering medical care to patients with NICH. It included several types of modern mathematical analysis, which allowed both to study a structure of NICH and to reveal and analyze factors, effecting an outcome of this severe pathology. It resulted in working out criteria of outcome prognostication and estimation of organization of NICH diagnosis.

Grouping tables were processed on a computer with the help of such programs, as Excel, Access, Statistica for Windows. It was followed by mathematical processing of the obtained data, using determination of standard statistical characteristics and an agreement criterion (c2), correlation, dispersion and regression analyses.

Results

The study showed, that the most frequent cause of NICH (in particular, in older patients) was hypertension in combination with atherosclerosis (75.81%). Aneurysm rupture was one of the main causes in patients under 40 (50.0%). Total mortality was 56.60%.

Multifactor analysis indicated, that the most significant factors, effecting NICH outcomes, was a degree of consciousness disturbances and severity of a state on admission, hematoma volume, a clinical-anatomic form of hemorrhage, a number of endogenous risk factors, a cause and number of episodes, age. The obtained regression equations, taking into account the above factors, make it possible to prognosticate NICH outcome in a certain patient.

Factors of prehospital and hospital stages turned out to effect an outcome. Reliable dependence of outcomes on prehospital diagnosis was typical of patients with clear consciousness. There was a tendency to increase of fatal outcomes in patients with moderate and severe torpor, being the result of misdiagnosis at a prehospital stage. It was discovered, that the smallest rate of mortality was watched in the above specialized neurosurgical department (37.37%); it reached 100% in other departments.

Mathematical analysis proved once again, that the most important diagnostic modalities in NICH were CT, angiography, lumbar puncture. Thus, it is expedient to use these methods of examination during the first hours, but not later than the first day since the moment of admission.

The results of our study indicate, that it is necessary to improve organization of medical care of patients with NICH. It implies improvement of diagnosis at a prehospital stage, admission to specialized hospitals, presence of up-to-date diagnostic apparatus (CT, MRI-CT), better knowledge of problems of vascular neurosurgery by doctors of outpatient departments and multi-field hospitals, specialists in emergency care.

Discussion

1. Clinical-Statistical Description of NICH

The most frequent cause of hemorrhage was hypertension, combined with atherosclerosis (75.8±1.6%). It is in compliance with data of literature [9, 10]. Among other causes of NICH one can mention aneurysm rupture (15.3±1.4%), tumor (1.5±0.5%), arterial hypertension (1.0±0.4%), AVM rupture (0.9±0.4%), atherosclerosis (0.2±0.2%). We failed to identify a hemorrhage cause in 5.4±0.9% of cases. According to other authors [2, 8], aneurysms, AVM, tumors account for more than a half of all NICH; the rest cases are caused by hypertension, blood diseases, infections, intoxication, etc. It appears, that this difference in causes and their quota in etiology of NICH can be explained by heterogeneous principles, used for solution of one and the same problem. Some authors carry out their studies under conditions of a neurologic department with a greater number of patients with NICH, caused by hypertension and atherosclerosis, others analyze data, obtained in neurosurgical departments with domination of cases with aneurysm and AVM ruptures.

The greatest number of intracranial hemorrhages was watched in people above 50 (76.1±1.6%). They developed only in 23.9±1.7% of cases under 50 (p<0.05). The most vulnerable age for men and women was 40-69 (78.7±2.4%; p<0.05) and 50-79 (72.3±2.3%; p<0.05) respectively.

The most frequent cause of hemorrhages in patients under 40 was aneurysm rupture (20-29 years old - 36.8±11.1%; 30-39 years old - 57.1±8.4%). Beligmotov B.Kh. [1], Deev A.S. and Zakharushkina I.V. [5] reported, that one of the main causes of intracranial hemorrhages in young people was vascular abnormalities. Hypertension, combined with atherosclerosis, occupied the first place in all age groups of patients above 40 (p<0.05). The majority of cases with an unclear cause of hemorrhages were 20-29 years old (52.6±11.5%). As the greatest number of hemorrhages in young people was usually conditioned by vascular defects, we stick to the opinion, that they were caused by microaneurysms and micromalformations, which were not diagnosed during examination in a hospital.

As for consciousness, the distribution of patients (682) was as follows: clear consciousness - 19.9±1.5%; moderate torpor - 26.5±1.7%; severe torpor - 14.8±1.4%; a soporific state - 18.8±1.5; coma I - 13.5±1.3; coma II - 6.0±0.9%; coma III - 0.4±0.3%. It demonstrates, that more than a half of cases had marked consciousness disorders on admission. The overwhelming majority of them was in a severe (49.9±1.9%) or extremely severe state (23.5±1.6%).

Many authors emphasize an important role of endogenous factors, including suffered and accompanying diseases, in development of NICH [11]. We studied and analyzed histories of all 682 patients. It was found out, that hypertension, combined with atherosclerosis (84.9±1.4%) and ischemic heart disease (62.9±1.9%) were the most frequent endogenous risk factor. The rest factors were as follows: stroke, suffered earlier - 14.8±1.4%; cardiac insufficiency - 7.9±1.0%; diabetes mellitus - 6.6±0.9%; obesity - 6.6±0.9%.

The analysis showed, that 248 out of 682 cases had intracranial hematomas, confirmed by CT findings, during an intervention or postmortem examination. All hematomas were localized supratentorially. Their volume was as follows: up to 30.0 ml - 39.9±1.0%; 31-50.0 ml - 10.9±1.0%; 51-100.0 ml - 27.8±1.0%; more than 100 ml - 21.4±1.0%.

Fatal outcomes were watched in 56.60 1.90% out of 682 patients (57.3±2.9% out of 300 males and 56.0±2.5% out of 382 females). A number of cases with fatal outcomes in male and female groups was approximately the same.

2. Multifactor Analysis of NICH Outcomes.

NICH Causes and Outcome. The greatest number of fatal outcomes was observed in patients with intracranial hemorrhage due to hypertension, combined with atherosclerosis, and aneurysm rupture (60.7±2.2% and 62.5±4.8% respectively, p<0.05) (Fig.1). There was no peculiar difference in dependence of fatal outcomes on a cause of hemorrhage among male and female patients (p>0.05).

Clinical and Anatomic Forms of NICH and Outcome. There was a direct proportional dependence of fatal outcomes on severity of a clinical-anatomic form of hemorrhage (R=0.98; p<0.05) (Fig.2). Fatal outcomes were watched in 20.3±3.6% of cases with subarachnoid hemorrhages; the same index in parenchymal-subarachnoid-ventricular hemorrhages was 88.2±2.4%. Almost analogous data were obtained by Shtuts V.N. [13]. According to his report, fatal outcomes were observed in 17.5 of patients with subarachnoid hemorrhages, 56.6% of cases with parenchymal or parenchymal-subarachnoid hemorrhages and 100% of patients with parenchymal-ventricular hemorrhage. As for our study, a lower index of mortality in patients with parenchymal hemorrhage and hemorrhage, accompanied by blood penetration into cerebral ventricles, is likely to be connected with improvement of diagnostic methods and treatment of this pathology.

Thus, the more severe a clinical-anatomic form, the higher a number of fatal outcomes.

Gusev V.I. et al. [4] concluded, that a number of fatal outcomes in patients with all types of stroke was directly dependent on severity of consciousness disturbance. We carried out analogous study in patients with intracranial hemorrhages and discovered, that fatal outcomes were watched in 23.5±3.6% out 136 patients with clear consciousness, 42.0±3.7% out of 181 cases with moderate torpor, 48.5±5.0% out of 101 patients with severe torpor, 78.9±3.6% out of 128 patients in a soporific state, 91.3±2.9% out of 92 cases with coma I. All the patients (100%) with coma II (42 men) and coma III (3 men) died.

These data indicate, that a growth of the rate of fatal outcomes is connected with aggravation of severity of consciousness disturbance. Besides, they demonstrate direct proportional dependence of an outcome on severity of this disturbance on admission (R=0.97, p<0.05) (Fig.3).

Mathematical analysis allowed to get regression equation (1), which can be used for approximate calculation of NICH outcome after determining a degree of severity of consciousness disturbance on admission.

(1)

Probability of a fatal outcome (%) = 17.865 + 13.311´С

State Severity and Outcome. A number of fatal outcomes in patients, admitted in a satisfactory state and a state of moderate severity was approximately identical (26.1±2.0% and 25.8±2.0% respectively, p=0.96). More than a half of patients, whose state on admission was severe, died (57.1±1.9%, p=0.0015). The greatest number of fatal outcomes (90.6±1.9%, p=0.005) was watched in patients, delivered in an extremely severe state. It is indicative of a growth of a number of fatal outcomes, when a general condition of a patient becomes more and more severe (R=0.94. p<0.05) (Fig.4).

Hematoma Volume and Outcome. Dependence of NICH outcomes on hematoma volume was studied. It was as follows: fatal outcomes were observed in 33.3±2.9% of patients with hematoma of 30.0 ml, 55.6±2.8% of cases with hematoma of 31.0-50.0 ml, 72.5±2.6% of patients with hematoma of 51.0-100.0 ml, 94.3±1.9% of cases with hematoma of more than 100.0 ml. Thus, there was a direct proportional dependence of NICH outcomes on hematoma volume (R=0.98, p<0.05) (Fig.5).

Risk Factors and Outcome. Some authors report a great effect of accompanying diseases on fatal outcomes in all types of stroke [3]. We revealed a direct proportional dependence of fatal outcomes on presence of accompanying or suffered pathology: the more risk factors, the worse an outcome (R=0.96, p=0.002). (p<0.05) (Fig.6). There was a growth of fatal outcomes from 36.8±3.9 (in patients with a single endogenous risk factor) up to 88.9±10.5% and 100.0% (in patients with 5 and 6 risk factors respectively).

Somatic Complications and Outcome.The most widespread somatic complication in patients with NICH was pneumonia. An effect of this complication on NICH outcomes was analyzed. Mortality in cases without pneumonia was 32.7±2.1%. The same index in patients with pneumonia reached 91.4±1.8%. The obtained data demonstrated, that pneumonia led to a threefold increase of mortality in patients with NICH and had a marked negative effect on prognosis.

Repeated Hemorrhages and Outcome. Many authors report a more serious outcome in repeated hemorrhages [1, 7]. Our study confirmed this information. Fatal outcomes were observed in 55.4±2.0% out of 589 patients with primary hemorrhage and 80.7±3.2% out of 93 cases with repeated hemorrhage.

Age and NICH Outcome.There was an extremely marked direct proportional dependence of outcomes on age (R=0.98, p<0.05): the older a patient, the greater a number of fatal outcomes. For example, this index was 21.1±9.4% in a group of 19 patients, aged 20-29, and 78.3±6.1% in a group of 46 cases of more than 80 years old (p<0.0001).

Mathematical Analysis of Factors, Effecting NICH Outcome. Mathematical processing of all available factors permitted to get linear regression equation and to describe their effect on mortality with the help of independent variables, designated as X (age, severity of consciousness disturbance, hematoma volume, a number of risk factors, etc.). This equation emphasizes once again significance of such parameters, as hematoma volume, age, a clinical-anatomic form of hemorrhage, severity of consciousness disturbance on admission, a number of risk factors. Substitution of these values into the obtained regression equation (2) allows to prognosticate probability of a fatal outcome in a certain patient.

(2)

Probability of a fatal outcome (%) = 11.4 +2.5 X1 +1.91 Х2+0.62 Х3+0.2 Х4+12.0 Х5

3. Factors of Prehospital and Hospital Stages and NICH Outcome.

Time of Admission and Outcome. Taking into account the fact, that literature contains no unanimous information on dependence of outcomes of acute vascular pathology on time of admission, we studied an effect of this factor on an outcome of intracranial hemorrhages. Time of admission is undoubtedly an important factor, conditioning an outcome. However, dispersion and regression analyses showed, that severity of consciousness on admission is a much more significant factor (F=53.94, p=0.0000001) in comparison with admission time (F=0.14, p=0.96). We failed to discover important difference in fatal outcomes in patients with analogous severity of consciousness disorders, who were delivered to a hospital during the first day, but at different time (hours) since the moment of hemorrhage.

Preshospital Diagnosis and Outcome.The analysis of prehospital diagnosis showed, that NICH was diagnosed more often by physicians of specialized neurologic emergency teams (79.5±4.4%). Doctors of ambulances (16.7±2.2%), outpatient departments (14.0±2.3) and doctor's assistants (3.0±3.0%) did it rather seldom. An effect of prehospital misdiagnosis on an outcome in patients with NICH was studied. There was a reliable dependence of outcomes on prehospital diagnosis in cases with clear consciousness (R=0.97). Fatal outcomes were watched in 11.8±5.5% of patients with clear consciousness and diagnosed NICH, this index was 2 (23.7±5.5%, p=0.17) and 3 (37.5±17.1%, p=0.09) times higher in patients with "diagnosed" acute disorder of cerebral circulation and ischemic stroke respectively. The same index in cases with other types of "diagnosed" pathology and diseases of other genesis was 7 (80.0±17.9%, p=0.001) and 7.5 (87.5±11.7%, p=0.001) times higher respectively. As for patients with moderate and severe torpor, there was a tendency to increase of a number of fatal outcomes, caused by misdiagnosis. However, we did not reveal any considerable dependence of outcomes on prehospital diagnosis in patients with more severe consciousness disturbances (a soporific state, coma I, II and III).

Dispersion analysis demonstrated, that severity of consciousness disturbance on admission (F=12.86, p<0.0001) was a more significant factor, conditioning an outcome, in comparison with prehospital diagnosis. The results of regression analysis (a step-by-step method) were indicative of synergism (interaction) of factors, i.e. both severity of consciousness disturbance and prehospital diagnosis effect NICH outcome, but in considerable dominance of the former. Regression equation (3) allows to carry out approximate estimation of NICH outcome on admission by taking into account severity of consciousness disturbance on admission and prehospital diagnosis.

(3)

Probability of a fatal outcome (%) = -15.99+19.1С + 14.7 DS - 2.79 C + DS

This equation can be used as a means of prognostication of NICH outcomes. It allows to carry out preliminary prognostication of outcomes of this severe pathology much quicker and, thus, promote rendering rapid and adequate aid in a great number of patients. The main prognostic sign, accelerating prognostication, is a degree of consciousness disturbance, being a more exact factor of estimating severity of a patient's state. The second prognostic sign is prehospital diagnosis, being expert estimation of a patient's state by a doctor at a prehospital stage.

A Hospital Level and Outcome.Today patients with NICH are admitted to hospitals with a different level of specialization and material and technical support. There are no reports, describing an effect of this ungrounded system of hospitalization on patients with NICH. Not a single study is devoted to a comparative description of NICH outcomes in patients, treated in hospitals of a different level. According to our opinion, only data of this type can substantiate and prove expediency of admission of patients with this severe pathology to specialized hospitals. Thus, we carried out detailed study and made comparative analysis of HICH outcomes in various department of hospitals of a different level. Indices of fatal outcomes were as follows: (1) a neurosurgical vascular department of the above-mentioned hospital of the third level - 37.4±3.4%, its neurologic department - 50.0±5.7%; (2) a neurosurgical department of the hospital of the second level - 63.6±8.4%, its neurologic department - 66.9±2.9%; (3) a neurologic department of the hospital of the first level - 66.4±4.6%, its cardiologic therapeutic and gastroenterologic departments, etc. - 100.0%. Outcomes of surgical treatment of NICH in the hospital of the third level were much better too; postoperative mortality was 40.5±5.7%. It was much higher in the hospital of the second level (71.4±8.5%). Thus, there was marked direct proportional dependence of outcomes on a hospital level (R=0.92, p=0.008).

Diagnostic Potentialities at a Hospital Stage and Outcome. The above data demonstrate an effect of a hospital level on mortality. As one of the main distinctive peculiarities of hospitals of different levels is existing diagnostic potentialities, we analyzed all methods of examination of patients with NICH, a term of their carrying out and their effect on outcomes. Regression analysis (a step-by-step method) confirmed once again, that the most significant diagnostic modalities, used in this pathology, were CT examination, angiography (AG), lumbar puncture (LP). It was discovered, that CT examination, AG and LP, applied respectively in more than 5, 3 and 5 days since admission, resulted in late diagnosis of NICH and increased mortality. Dependence of fatal outcomes on a term of carrying out these examinations can be expressed by the following equation (4):

(4)

Mortality = 102.1 - 1.31 AG3 - 11.75 LP5 + 11.56 LP3 - 0.19 CT5

Determination of a term of carrying out the above examinations can be followed by prognostication of mortality, caused by NICH, in departments of different hospitals. It is of great importance for solving problems of organization and management, analysis of hospital activity and control of use of diagnostic devices.

Conclusion

1. Hypertension, combined with atherosclerosis, was the most frequent cause of NICH (75.80% of patients). Thus, it is necessary to carry out screening in outpatient departments and at big enterprises with the purpose of revealing hypertension and atherosclerosis, being the main risk factors of NICH.

2. Complex mathematical analysis demonstrated, that the most significant factors, effecting NICH outcome, were a degree of consciousness disturbance and severity of a state on admission, hematoma volume, a clinical-anatomic form of hemorrhage, a number of endogenous risk factors, a cause and number of hemorrhages, age of a patient. The above regression equations, taking into account all these factors, can be used for prognostication of NICH outcome in a certain patient.

3. There is a statistically reliable dependence of outcomes on prehospital diagnosis in patients with clear consciousness. As for cases with moderate and severe torpor, there is a tendency to a growth of fatal outcomes, resulting from prehospital misdiagnosis. The first medical aid should be rendered by doctors of specialized neurologic teams. It is necessary to improve professional skills of doctors and doctor's assistants of mobile emergency teams, therapeutists and neurologists of outpatient departments in the field of NICH diagnosis.

4. Dispersion and regression analyses confirmed once again the fact, that the most useful diagnostic modalities at a hospital stage were CT, angiography and lumbar puncture. Application of CT, AG and LP in more than 5, 3 and 5 days since admission respectively resulted in a growth of mortality. Thus, the optimum tactics lies in use of these methods of examination during the first hours and no later, than the first day since the moment of NICH.

5. Mortality in the specialized neurosurgical department and departments, specializing in other fields of medicine, was 37.37% and 100.0% respectively; as for postoperative mortality it was equal to 40.5% and 71.4%. It proves, that patients with suspected NICH should be admitted to specialized hospitals, which have highly qualified neurosurgeons, possessing a good knowledge of this pathology.

REFERENCES

  1. Belimgotov B.Kh. A system of specialized medical care in the Kabardin-Balkar Republic and treatment of aneurysmal intracranial hemorrhages: The thesis for a degree of a Doctor of Medical Science.-Saint Petersburg, 1998.-376 P. (Rus.).
  2. Belimgotov B.Kh., Kozhaev Z.U. An acute period of a rupture of intracranial aneurysms: A system of organization, diagnosis and tactics at a prehospital stage//Problemy neirochirurgii: A collection of scientific reports.-Saint Petersburg, 2000.-P. 121-122 (Rus.).
  3. Vilensky B.S., Semyonova G. Causes of death in stroke and possible measures of decreasing mortality//Nevrologichesky Zhurnal.-2000.-N 4.-P. 10-13 (Rus.).
  4. Gusev E.I., Vilensky B.S., Skoromets A.A. et al. Main factors, effecting stroke outcomes//Zhurnal nevropatologii i psikhiatrii.-1995.-V. 95, N 1.-P. 4-7 (Rus.).
  5. Deev A.S., Zakharushkina I.V. Cerebral strokes in young people// Zhurnal nevropatologii i psikhiatrii.-2000.- N 1.-P. 14-17 (Rus.).
  6. Kalushev Yu.N., Longinidi A.A. Hemorrhagic strokes - peculiarities of modern tactics of treatment under conditions of a city neurologic hospital//Actual. problemy nevrologii i neirokhirurgii; A collection of scientific reports.-Rostov-on-Don, 1999.-P. 103-104 (Rus.).
  7. Lebedev V.V., Krylov V.V. Emergency neurosurgery: A manual for doctors.-Moscow, 2000.- 568 P. (Rus.).
  8. Nikiforov B.M., Zakaryavichus Zh. A hemorrhagic period of cerebral aneurysms: Results of multimodality treatment//Problemy neirokhirurgii: A collection of scientific reports.-Saint Petersburg, 2000.-P. 141-143 (Rus.).
  9. Odinak M.M. et al. Vascular diseases of the brain.-Saint Petersburg, 1998.-P. 99-116 (Rus.).
  10. Polishchuk N.E., Rasskazov S.Yu. Principles of treatment of a patient in emergency neurology and neurosurgery.-Kiev, 1998.-P. 60-63 (Rus.).
  11. Kharakoz O.S. et al. Arterial hypertension as a main risk factor of cerebral stroke// Kardiologiya.-200.-N 1.-P. 43-47 (Rus.).
  12. Shmatko V.G., Bagir V.N. A clinical picture of hemorrhagic stroke//Emergency neurology: Transaction of the Research and Practice Conference.-Omsk, 1998.-P. 105-106 (Rus.).
  13. Shtuts V.N. Diagnosis and treatment of hemorrhagic stroke and specialized emergency care: Abstract of thesis for a degree of a Candidate of Medical Science.-Sverdlovsk, 1968 (Rus.).
  14. Ozava T. et al. Primary angiitis of the central nervous system: report of two cases and review of the literature//Neurosurgery.-1995.-Vol. 36, N 1.-P. 173-179.